*** grantc has joined #ingres | 00:10 | |
*** Mud has joined #ingres | 00:16 | |
Deyan | hiya | 00:26 |
---|---|---|
*** stliu has quit IRC | 01:52 | |
*** Alex| has quit IRC | 02:31 | |
*** Alex| has joined #ingres | 03:14 | |
*** ChanServ sets mode: +o Alex| | 03:14 | |
*** KermitTheFragger has joined #ingres | 03:18 | |
*** Alex| has quit IRC | 03:44 | |
*** rossand has joined #ingres | 04:26 | |
*** ChanServ sets mode: +o rossand | 04:26 | |
Vroomfondle | grantc: had any reports of problems with CREATE queries via PHP driver (specifically, warnings about being unable to close the statement)? Am trying to troubleshoot a long-running problem we've had which is causing (apparently perfectly legal) CREATEs to fail. | 04:29 |
Vroomfondle | ..? | 04:29 |
Vroomfondle | I can't see anything in the bug tracker which matches | 04:29 |
Vroomfondle | the warning is preceded by a couple of "Server or API error ... status 7" warnings from ii_success() | 04:30 |
* Vroomfondle is ploughing through ingres.c but isn't much good at C debugging | 04:31 | |
Deyan | -rw-r--r-- 1 ingres ingres 58G Nov 22 09:09 errlog.log | 04:33 |
Deyan | LOL | 04:33 |
grantc | Vroomfondle: this might be a fixed problem | 04:38 |
grantc | let me see if I can find the bug/change | 04:38 |
grantc | http://pecl.php.net/bugs/bug.php?id=18006 | 04:40 |
grantc | do you execute the create without assigning the return value from ingres_query()? | 04:41 |
grantc | i.e. "ingres_query(...)" vs "$rs=ingres_query(..)" | 04:42 |
Vroomfondle | no - all calls to ingres_query go through a wrapper, and the wrapper always assigns to a resultset | 04:56 |
Vroomfondle | grantc: see http://pastebin.com/DmcUvsky for a full list of the warnings generated by a single CREATE | 04:57 |
* Vroomfondle wanders off for lunch - BBL | 04:57 | |
Vroomfondle | (oh, btw, the wrapper is literally just doing "$results = ingres_query(..." and then later on $results will fall out of scope and be garbage-collected | 04:58 |
Vroomfondle | autocommit is off | 04:58 |
Vroomfondle | (but we fake our own autocommit) | 04:58 |
* Vroomfondle really goes to lunch now, honest | 04:59 | |
grantc | i'm guessing a test case it out of the question? | 05:00 |
*** cthibert has joined #ingres | 05:13 | |
Vroomfondle | grantc: unfortunately, this is the test case: | 05:28 |
Vroomfondle | CREATE TABLE test (test char(1)) | 05:28 |
Vroomfondle | (or any other CREATE TABLE as far as I can tell so far) | 05:28 |
Vroomfondle | I can furnish you with various details about our systems of course, if you let me know what you need | 05:28 |
*** stliu has joined #ingres | 05:29 | |
Vroomfondle | there are other oddities, such as the fact that although I'm fairly sure it only used to crash on our development server, now it seems to crash when run on our "demo" (user testing) system as well. The dev system is Ingres 9.2 running on Sol10, and the demo system is on Ingres 2.6 running on Sol9 (according to our sysadmin). | 05:32 |
Vroomfondle | we're using driver 2.2.3 on dev and 2.0.2 on demo | 05:32 |
Vroomfondle | so I'm really struggling to find any commonalities other than the application code, but as far as I know everything we're doing is just standard stuff (I wrote the wrapper code) | 05:33 |
*** cthibert has left #ingres | 05:35 | |
Vroomfondle | ooh, hang on | 05:36 |
*** cthibert has joined #ingres | 05:37 | |
Vroomfondle | ...no, damn :( | 05:38 |
Vroomfondle | aha! | 05:40 |
Vroomfondle | it occurs when I try to iterate over the result-set after doing a CREATE | 05:40 |
Vroomfondle | naturally there won't be any results, but I'd assumed it was safe to just do that for every query | 05:40 |
Vroomfondle | grantc: is that a bug, or am I just Doing It Wrong? | 05:40 |
Vroomfondle | (I'm doing: while($results_handle && ($row = ingres_fetch_array($results_handle))) | 05:41 |
Vroomfondle | ) | 05:41 |
Vroomfondle | etc. | 05:41 |
grantc | 18006 is a bug - the un-assigned result handle gets garbage collected (closing resources etc) and then on the next query the driver tries to do the same | 05:42 |
grantc | it could be that the fix for 18006 also helps in this situation | 05:43 |
Vroomfondle | has the fix made it into an actual release yet, or shall I load up an SVN client? | 05:44 |
grantc | only in svn right now | 05:49 |
* grantc wanders off to prepare lunch | 05:49 | |
Vroomfondle | okay, cheers | 05:51 |
*** mull has quit IRC | 06:00 | |
*** atrofast has joined #ingres | 06:15 | |
*** atrofast has joined #ingres | 06:16 | |
Vroomfondle | grantc: just a bit of further info - this appears to be the same issue which is sometimes causing segfaults when we do INSERTs on our dev system. Anyway I'll see about getting the SVN version running, and stop pestering you :D | 06:28 |
*** DarylM has joined #ingres | 06:29 | |
*** mull has joined #ingres | 06:42 | |
Deyan | wow, some activity here :) | 06:49 |
Deyan | i did not see this bug | 06:49 |
Deyan | maybe because I build svn quite often | 06:49 |
grantc | Vroomfondle: that's good to hear :) | 07:28 |
grantc | Deyan: i'm surprised you've not hit this as it's probably been there for a while | 07:29 |
*** cthibert has left #ingres | 08:00 | |
*** cthibert has joined #ingres | 08:02 | |
*** cthibert has left #ingres | 08:13 | |
*** cthibert has joined #ingres | 08:13 | |
*** stliu has quit IRC | 08:39 | |
Deyan | i will check it | 09:48 |
Deyan | time to go home guys | 09:48 |
Deyan | see you tomorrow | 09:48 |
*** Deyan has quit IRC | 09:48 | |
*** grantc has quit IRC | 09:53 | |
*** Alex| has joined #ingres | 10:36 | |
*** ChanServ sets mode: +o Alex| | 10:36 | |
*** monda07 has joined #ingres | 11:13 | |
*** monda07 has quit IRC | 11:58 | |
*** KermitTheFragger has quit IRC | 12:09 | |
*** rossand has quit IRC | 12:10 | |
*** rossand has joined #ingres | 12:11 | |
*** ChanServ sets mode: +o rossand | 12:11 | |
*** atrofast has quit IRC | 13:01 | |
*** cthibert has left #ingres | 13:04 | |
*** Mud has quit IRC | 13:46 | |
*** Alex| has quit IRC | 13:48 | |
*** spectrahp has joined #ingres | 14:14 | |
*** spectrahp has left #ingres | 14:17 | |
*** mull has quit IRC | 15:37 | |
*** mull has joined #ingres | 16:21 | |
*** rossand has quit IRC | 17:21 | |
*** stliu has joined #ingres | 18:13 | |
*** stliu has joined #ingres | 18:13 | |
*** cytrinox_ has joined #ingres | 21:42 | |
*** cytrinox has quit IRC | 21:46 | |
*** cytrinox_ is now known as cytrinox | 21:46 |
Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!